Are you guilty of irrational behaviour when the petrol price goes up?

So petrol goes up 32c per litre tomorrow (7 August). When I was a child I never understood why my parents seemed obsessed with the price of petrol. Now, as an adult, my eyes water every time I see the cost of filling my tank up.

Motorist behaviour around the time that petrol prices increase has always interested me. In particular, I have never quite understood why people insist on going out of their way to fill up their tank the day before petrol goes up. It might seem obvious as to why they do this, but consider this angle.

A while back I was at a friend’s house the night before the petrol price was due to go up by around 30c. They remembered that the price was due to go up the next day, so they promptly got in their car and drove off to the nearest petrol station to fill up. It was an inconvenience, but my friend was happy that they had got one over the petrol station.

This is not an uncommon event, but here’s why I don’t understand it:

  • Firstly, if you assume that you only buy 40 litres of petrol (remember, your tank isn’t empty), you’re only saving R12 (40c x 30). 
  • Secondly, and more importantly, if my friend was at home at 9pm at night and I offered them R12 to drive to the petrol station and back, they probably wouldn’t do it because R12 is not worth the hassle.

And that’s where the irrationality comes in. We should be indifferent between saving R12 by buying petrol, or being given R12 for going through the hassle involved in filling up. But for some reason, a lot of us are not. A large number of people will happily go through the inconvenience of filling up now, but they wouldn’t go through the same hassle if you paid them the same amount of money.

This post was just intended to be a light-hearted, topical piece around consumer irrationality. But, this subject is more important than you might think. Through designing insurance products, I have learnt that policyholders and financial advisers often make irrational decisions around their risk cover. I won’t try to explain why this is so, but I will remind you that when making financial decisions, it is important to to take emotion out of your decision making process as far as possible, as well as constantly making sure you are able to justify your decisions to yourself.

And if you are going to rush off to the petrol station tonight, at least make it worth it and get an ice cream at the same time.


The impact of telematics on your car insurance premiums: A pricing actuary’s thoughts

In the last year or so, telematics has become the new buzzword in car insurance. Most of the noise I have heard is that the introduction of telematics into the car insurance industry will result in lower insurance premiums for those who drive responsibly, and higher premiums for those who drive recklessly. I have my doubts that this is a perfect solution for determining insurance premiums though. Let me explain.

1. What is telematics?

If you opt for the telematics insurance option, a telematics device will be installed in your car. This device then monitors your driving behaviour and reports this back to your insurer. The device can monitor your acceleration,  braking, cornering, speed and numerous other aspects of your driving. The more responsibly you drive, the lower the likelihood that you will have a claim on your car insurance and the lower the premium the insurer needs to charge you. If you drive recklessly, the opposite applies. The thinking is that telematics will mean that the insurer no longer needs to charge you a premium that represents the average level of driver competency, but rather the premium that reflects your specific competency as a driver.

2. What are the potential shortfalls  of telematics?

I have three main concerns about how telematics will be used in the car insurance industry. These are:

  • The aspects that telematics monitor are not a perfect indicator of your claims likelihood

While the theory behind telematics sounds great, I’m not convinced that your telematics score is the perfect indicator of your liklihood of having a claim. My grandmother would probably be classed as a low risk by telematics. She doesn’t speed, doesn’t corner aggressively and she doesn’t accelerate quickly. However, being in the same car as her is a terrifying experience, because she is just a terrible driver. On the flip side, I know a number of excellent drivers who do drive recklessly according to any telematics evaluation, but I would sooner get in the car with them than my grandmother.

A particular example is speed. Say one telematics device records one driver regularly travelling at 130 km/h, and another travelling at 100 km/h. At face value, the former is higher risk because he is definitely speeding. However, I would argue that someone travelling at 130 km/h on a four lane freeway with a speed limit of 120 km/h is a lot lower risk than someone travelling at 100 km/h in a residential area with a speed limit of 60 km/h.

  • The difficulty of turning all the data provided by telematics into a factor to load premiums by

The data provided by telematics is by no means simple. Whoever decides on the formula that decides how telematics data impacts on your premiums will need to decide how to weight the myriad of different aspects that the telematics device reports to the insurer.

More importantly, whoever does this will likely not have enough historic data on which to decide which factors the telematics device reports on have the biggest impact on your claim likelihood, since the telematics device has only recently become affordable and used widely. In order to do this properly, copious amounts of data is required. This makes an already difficult task quite subjective, requiring the person who determines how premiums are impacted by telematics data to use gut feel and how they think the telematics data should be interpreted, instead of objective historic data.

  • There are other factors that already used that determine your risk profile

I don’t price car insurance, but I would imagine the main factors that influence your car insurance premiums (from an accident perspective, not theft) are car model, your monthly mileage, your age and your gender.  If telematics is to be used correctly, it needs to factor in these aspects before impacting your premium.

Confused? Let’s consider a simple example. Young people tend to pay more for car insurance, partly because they are more reckless drivers. Before telematics data is used to influence the premiums for a young life, they already pay  a higher premium because young people in general are more reckless. Telematics only adds value to the extent that it measures how reckless you are compared to the average person similar to you, not the average person in general.

If telematics does not take this into account and just adjusts premiums compared to the average driver overall, then the groups who are generally more risky will get double damned, and those who already drive better in general will be given too big a premium reduction. Again, the only way to correctly allow for telematics is to have very large amounts of telematics data for all types of drivers, together with their claims history, which is not available yet.

3. Conclusion

From the above you’ll  gather that I don’t believe that telematics is the answer to providing every policyholder with an accurate premium for the way that they specifically drive. However, I am all for telematics being introduced in car insurance. In South Africa in particular with our horrific accident rate, any development that reimburses drivers for driving less aggressively (even if it’s only on the telematics score) can only be a good thing.